Samy A. Mahmoud

Optimal allocation of resources in distributed information networks

The problems of file allocation and capacity assignment in a fixed topology distributed computer network are examined. These two aspects of the design are tightly coupled by means of an average message delay constraint. The objective is to allocate copies of information files to network nodes and capacities to network links so that a minimum cost is achieved subject to network delay and file availability constraints. A model for solving the problem is formulated and the resulting optimization problem is shown to fall into a class of nonlinear integer programming problems. Deterministic techniques for solving this class of problems are computationally cumbersome, even for small size problems. A new heuristic algorithm is developed, which is based on a decomposition technique that greatly reduces the computational complexity of the problem. Numerical results for a variety of network configurations indicate that the heuristic algorithm, while not theoretically convergent, yields practicable low cost solutions with substantial savings in computer processing time and storage requirements. Moreover, it is shown that this algorithm is capable of solving realistic network problems whose solutions using deterministic techniques are computationally intractable.

Robust voice activity detection using higher-order statistics in the LPC residual domain

This paper presents a robust algorithm for voice activity detection (VAD) based on newly established properties of the higher order statistics (HOS) of speech. Analytical expressions for the third and fourth-order cumulants of the LPC residual of short-term speech are derived assuming a sinusoidal model. The flat spectral feature of this residual results in distinct characteristics for these cumulants in terms of phase, periodicity and harmonic content and yields closed-form expressions for the skewness and kurtosis. Important properties about these cumulants and their similarity with the autocorrelation function are revealed from this exploratory part. They show that the HOS of speech are sufficiently distinct from those of Gaussian noise and can be used as a basis for speech detection. Their immunity to Gaussian noise makes them particularly useful in algorithms designed for low SNR environments. The proposed VAD algorithm combines HOS metrics with second-order measures, such as SNR and LPC prediction error, to classify speech and noise frames. The variance of the HOS estimators is quantified and used to yield a likelihood measure for noise frames. Moreover, a voicing condition for speech frames is derived based on the relation between the skewness and kurtosis of voiced speech. The performance of the algorithm is compared to the ITU-T G.729B VAD in various noise conditions, and quantified using the probability of correct and false classifications. The results show that the proposed algorithm has an overall better performance than G.729B, with noticeable improvement in Gaussian-like noises, such as street and parking garage, and moderate to low SNR.

Efficient search and design procedures for robust multi-stage VQ of LPC parameters for 4 kb/s speech coding

A tree-searched multistage vector quantization (VQ) scheme for linear prediction coding (LPC) parameters which achieves spectral distortion lower than 1 dB with low complexity and good robustness using rates as low as 22 b/frame is presented. The M-L search is used, and it is shown that it achieves performance close to that of the optimal search for a relatively small M. A joint codebook design strategy for multistage VQ which improves convergence speed and the VQ performance measures is presented. The best performance/complexity tradeoffs are obtained with relatively small size codebooks cascaded in a 3-6 stage configuration. It is shown experimentally that as the number of stages is increased above the optimal performance/complexity tradeoff, the quantizer robustness and outlier performance can be improved at the expense of a slight increase in rate. Results for log area ratio (LAR) and line spectral pairs (LSPs) parameters are presented. A training technique that reduces outliers at the expense of a slight average performance degradation is introduced. The method significantly outperforms the split codebook approach. >

A comparison of trellis coded versus convolutionally coded spread-spectrum multiple-access systems

A system model is proposed that allows one to apply both trellis coding and PN spreading sequence to the data symbols to be transmitted. Rate n/n+1, trellis codes using 2/sup n+1/-point MPSK signal constellations are investigated when Gold sequences are used for PN spreading. Performance in an additive white Gaussian noise (AWGN) channel is investigated, with 5-20 users transmitting simultaneously. Using the criteria of equal complexity and throughput, the performance of the trellis codes in a SSMA (spread spectrum multiple access) environment is compared to that of medium-rate to low-rate convolutional codes through the use of a generalized transfer function bound. The average degradation due to the interuser interference is determined by the method of moments. The validity of approximating the interuser interference as a Gaussian random variable is also investigated. The numerical results illustrate that for a given complexity, chip rate and throughput, low-rate convolutional codes provide the best performance in an SSMA system. As lower-rate convolutional codes are used, there is an increase in the effective interuser interference due to the greater cross-correlation effects from using shorter PN sequences, or alternatively from the effects of partial cross-correlation. However, this increased degradation is more than overcome by the increased distance properties of the low-rate codes. >

Selection diversity combining with multiple antennas for MM-wave indoor wireless channels

Presents predetection and postdetection combining schemes for selection diversity reception with multiple antennas for MM-wave indoor radio channels. For those combining schemes, a reduction in complexity is achieved by limiting the number of combined signals to small values and by increasing the number of received signals. Bit error rate (BER) performance of binary phase shift keying (BPSK) with predetection combining of selected signals (CSS) and BER performance of differential BPSK with postdetection CSS are analyzed for slow fading and Rayleigh-distributed envelope statistics. Predetection maximal ratio combining of signals that comes from a single group or several groups of diversity channels as well as postdetection combining of received signals for groups of channels are considered. In comparing predetection combining with groups (PCG) and predetection combining of the best signals (PCB), we observe that the required SNR for achieving a certain BER is approximatively the same (with PCG having a slight advantage of 0.5 dB) for a given number, N, of diversity channels and L combined signals. Furthermore. PCG is equivalent to PCB for L=N since both techniques then correspond to conventional predetection maximal ratio combining (MRC), PCG and PCB are also equivalent when L=1 as both schemes then correspond to conventional selection combining. A small degradation of approximately 2 dB in the required SNR is observed when postdetection diversity reception with groups (PDG) is used instead of PCG. For L=N, PDG reduces to post detection MRC. The PDG technique is considered more suitable than PCB or PCG for MM-indoor wireless systems.

Compound strategies of coding, equalization, and space diversity for wide-band TDMA indoor wireless channels

Compound strategies of equalization and space diversity in the form of an optimum baseband combiner are attractive for wideband time division multiple access (TDMA) portable communication radio links in order to combat dispersive fading and cochannel interference. The authors investigate the performance of such a scheme in conjunction with convolutional coding and soft-decision Viterbi decoding via a semianalytical technique based on the method of moments. Such an approach avoids a Gaussian characterization of interference and yields results for both ideal interleaving and no interleaving. With dual space diversity, three taps per forward filter, and a data rate of 10 Mb/s, it is shown that, although a third space diversity branch remains preferable in terms of performance, channel coding can be a viable alternative, particularly in terms of outage rate, to increasing the space diversity order, even in the absence of interleaving, provided the signal-to-interference ratio is sufficiently high. >

A link between quasi-cyclic codes and convolutional codes

Given a quasi-cyclic code with minimum Hamming distance d, a set of convolutional codes is derived with free distance equal to d. It is shown that an increase in the rate of these codes results in a decrease in the memory length. The connection between these codes is illustrated with several well-known quasi-cyclic codes. The free distance of some partial unit memory convolutional codes can be determined using the results in this correspondence.

Filtered decision feedback channel estimation for OFDM-based DTV terrestrial broadcasting system

Orthogonal frequency division multiplexing (OFDM) is one of the transmission techniques used for digital television (DTV) terrestrial broadcasting. A high quality channel estimator with a low training overhead is the key to the successful delivery of DTV services which require high spectrum efficiency and robustness to strong and dynamic ghosts. Robustness to multipath distortion is especially important in a single frequency emission environment. This paper presents an application of a filtered decision feedback channel estimator for OFDM-based DTV systems using high order QAM modulations. The implementation and the performance of the channel estimator are discussed. Computer simulations were conducted to evaluate the performance of the channel estimator. The channel estimation loss is about 1.2 dB from the ideal case where the channel is assumed to be known by the receiver. For a given multipath spread, the loss can be further reduced by increasing the FFT size. The FFT size is however subject to an upper limit imposed by the Doppler spread.

Hardware nonlinearities in digital TV broadcasting using OFDM modulation

FFT-based coded orthogonal frequency division multiplexing (COFDM) is one of the techniques for digital TV broadcasting over multipath fading channels. A FFT-based OFDM signal is subject to various hardware nonlinearities in both the transmitter and receiver. Hardware nonlinearities not only affect the in-band performance of an FFT-based OFDM system but also may affect the system performance of an adjacent channel signal because of regenerated sidelobes of the transmitted signal. The paper investigates the in-band and out-of-band behaviour of a 64QAM-OFDM system under various nonlinear devices. It is shown that the inherent signal clipping in the IFFT processors with a limited word length reduces the required RF amplifier output backoff (OBO) where adjacent channel interference is the limiting factor. For a 0.25% clipping rate, an additional 2 dB OBO is required for the COFDM signal to achieve the same level of adjacent channel interference as for the single carrier system. The loss in SNR due to signal clipping is negligible in a coded OFDM system.

SNR estimation of speech signals using subbands and fourth-order statistics

This article addresses the problem of instantaneous signal-to-noise ratio (SNR) estimation during speech activity for the purpose of improving the performance of speech enhancement algorithms. It is shown that the kurtosis of noisy speech may be used to individually estimate speech and noise energies when speech is divided into narrow bands. Based on this concept, a novel method is proposed to continuously estimate the SNR across the frequency bands without the need for a speech detector. The derivations are based on a sinusoidal model for speech and a Gaussian assumption about the noise. Experimental results using recorded speech and noise show that the model and the derivations are valid, though not entirely accurate across the whole spectrum; it is also found that many noise types encountered in mobile telephony are not far from Gaussianity as far as higher statistics are concerned, making this scheme quite effective.